Accelerated Design of Catalytic Water-Cleaning Nanomotors via Machine Learning.

The ability of self-propelled nanoparticles to convert environmental energy into locomotion has led to several nanomotor prototypes that are promising in numerous real-world applications. However, the vast variety of nanoparticle designs prevents rapid identification of the optimal composition for a given application. In this study, we applied machine learning methods to predict the self-propulsion speed and water-cleaning efficiency of micro/nanomotors (MNMs), where the quality of machine learning predictions was evaluated based on the statistical values. The average absolute error of predicted velocity and predicted efficiency are determined to be as low as 0.10 and 0.12, respectively. In addition, by comparing the prediction results based on 13 features using four different machine learning algorithms, we are able to identify several key features that are important to effectively environmental decontamination, such as particle size, catalyst type, and aspect ratio. Following the guidelines deduced from these models, a high-efficiency Pt-coated nanomotor was designed and synthesized, of which the experimental results were compared with the machine learning predictions, showing an accurate prediction with a less than 15% of prediction error. In the range of our theoretical/experimental conditions, we showed that a gradient boosting algorithm is the most promising method for predicting the environmental decontamination behavior of MNMs, a machine-learning algorithm rarely used in the nanomaterial field in current practice.